[[TracNav(TracNav/ISO15926Primer)]] = We Need to Focus on Business Processes = Status of this document: Ready for Cold Eyes Review ---- [[PageOutline(2-4,Contents,inline)]] == Abstract == The barrier to interoperability between software applications is not technology. The barrier is business processes. We typically approach interoperability with more technology. Instead we should start with information modeling. ---- == Traditional Approach: Focus on Technology == Every organization today is under relentless pressure to reduce costs. One way to reduce costs is to increase productivity, and nowadays, this is usually accomplished by automating manual tasks by writing some computer application. [[Image(BusinessProcess01.JPG, 300px)]] ''' Fig 1 - Problem Space ''' Plant design nowadays involves work processes that (more-or-less) take information from one or more software applications, transform it somehow with a manual process, and enter the results into another application. These work processes may range from precise, documented procedures to heuristics and rules of thumb that "people are just supposed to know". In most organizations there are hundreds of them. === Example: Populate a PO Directly From 3D Model Database === There are several 3D plant modeling applications available commercially. There are also several purchasing applications available commercially. But none of the purchasing applications, out of the box, will populate a purchase order directly from any of the 3D modelling applications. But building purchase orders based on the contents of a 3D model is something every EPC has to do a great many times. In order to do this, every organization has had to develop work processes for selecting a range of objects from a 3D model, extracting a material report in some format, transforming the report into something the purchasing application can understand, then importing it to the purchasing application. If this sort of thing has to be done often enough, there is an opportunity to increase productivity and reduce errors by automating it. Using today's technology there are a great many ways to do the automation: * Commercial Middleware * Custom programming (C, Java, Visual Basic, ...) * Neutral file (Excel, Notepad) * XML Neutral file We can make this work because, for any two particular applications, we know the context. So it's easy for the developer to hard code the context into the bridging application, often simply ignoring it if she knows the meaning of the respective data objects is the same. This works for a closed system, but will not work in a world where software has to interoperate automatically, with no prior knowledge of each other's data structure. === What is the Problem? === The problem here is similar to that of point-to-point mapping, of which we have written at length elsewhere in this Primer. * The costs of setup and configuration * Repeating costs for setup and configuration * The cost of interpretation (because meaning is often inferred from the context) * The complexity is exposed (you have to do it from scratch each time) * The new global, distributed work execution model amplifies the problem * Aging Workforce (these are the people who understand the context of legacy systems) == Alternative: Focus on Information Modeling == [[Image(BusinessProcess02.JPG, 300px)]] ''' Fig 2 - A New Approach''' The traditional approach is to automate an existing work process. The better way is to understand the information model and focus back on the business task we want to accomplish. In the purchasing order example above, the task is to get descriptions of certain components from a 3D model into a purchase order. If we solve the problem by precisely modeling the information we get from the 3D model in a way that preserves the meaning (i.e., the semantics) of the data values, we will be able to use the solution more than once. It will take longer to model the information in this way, but we will only have to do it once, even if the software vendor publishes a new version, or even if we want to use a different set of software on a new project. Admittedly, the statement "precisely modeling the information we get from the 3D model in a way that preserves the meaning of the data values" is not a trivial statement. It's sort of like saying "It's ''easy'' for pigs to fly, all they have to do is ''learn how to grow wings''." When you get under the hood with information modeling, it starts to look like Artificial Intelligence. After all, we want a machine to be able to read any random data sheet and understand the context--the meaning--just as well as a human being. [[Image(BusinessProcesses_FlyingPig.JPG, 100px)]] Fortunately, with ISO 15926 this pig can indeed learn how to grow wings. === ISO 15926 is a Robust Standard === If an organization wanted to model its information it would not have to use ISO 15926. For instance, it could create its own standard. But the semantic precision that is required to be interoperable is elusive. It would need the ability to have correct grammar, and would need a mapping methodology to map business objects into the precision. ISO 15926 brings all that with it. So there are considerable advantages to using ISO 15926: * You leverage the work of a great number of people, from many industries world wide. * You do not have to design the system of recording knowledge. * You do not have to design every entity yourself, you can use entities designed by others, using some objects verbatim if appropriate, or extend existing objects and only have to add the differences. * The resulting applications will interoperate with other ISO 15926-enabled applications. Information Modeling is not an exact science. But the standardized methodology of ISO 15926 will ensure that two information models will align even if they originate at different organizations that have no knowledge of each other. The templates and classes of ISO 15926 are public and extensible, with the extensions (optionally) being public as well. The last section of this Primer, [wiki:ISO15926Primer_GettingStarted Getting Started with ISO 15926], will discuss some resources for learning Information Modeling and strategies for getting involved in ISO 15926. == Next == [wiki:ISO15926Primer_Benefits How ISO 15926 Will Make Your Life Easier] There are a number of issues associated with the need for interoperability, or with current attempts at interoperability. This section shows how ISO 15926 will help. === Acknowledgements === Thanks to Robin Benjamins for Figures 1 and 2. ----